RTF analysts developed per unit technical potential estimates for demand response impacts in residential water heaters. Both electric resistance and heat pump water heaters were studied. The RTF approved the analysis methodology at the June 2019 meeting.
Demand response impacts from water heaters occur by reducing or eliminating the use of electricity for water heating during event hours. Theoretically, this can be accomplished using one or a combination of methodologies, including pre-heating water, adjusting set points, preventing the use of electric resistance heating elements in HPWHs, using mixing values, or interrupting power to all heating elements. Generally, electricity used for water heating is shifted to a different time of the day, but in some cases, aggregate energy usage may be reduced if a household uses water heated to a colder temperature during an event than otherwise would have been the case.
RTF Presentation Residential Hot Water Heaters DR Workbook DR Subcommittee Presentation DR Subcommittee Notes
Approach
RTF analysts utilized existing demand response studies of residential water heaters to estimate potential per unit impacts under a range of scenarios. This approach was due to the inability to utilize the RTF version of SEEM to reliably model DR impacts from water heaters. To arrive at potential per unit impacts, analysts strove to adjust evaluation results in order to backout impacts due to program design and implementation, such as participant opt-outs and connectivity issues. DR impacts were estimated at the hourly level for all hours of the day and months of the year. Additionally, impact estimates were made for shed events, where customer water heaters are able to override DR event signals if additional hot water is needed to meet household demands, and emergency events, where water heating is not available during an event.
The estimates are primarily based on a recent regional study of CTA-2045 enabled water heaters lead by BPA. However, there are legitimate concerns that the participant sample in this study may not be indicative of typical DR program participants. Nonetheless, the impact results were reasonably corroborated by other studies reviewed by RTF analysts for this effort, including lab work by PNNL and findings from other jurisdictions.
Assumptions
All evaluations and studies used in the RTF analysis were focused upon hours of system peak demand, typically morning and evening hours in the summer or winter, depending on the jurisdiction. While these are likely to be the primary hours of interest to the region as well, the RTF analysis attempts to utilize load shapes to estimate the potential DR impacts for events called in other hours of the day and months of the year. This portion of the analysis is based on the assumption that the proportional magnitude of the assumed load curve for end use water heating is directly related to the proportional DR impact for an average household in any given hour of the year. This assumption is likely only partially true, but serves as a beginning point for the purposes of estimation.
Limitations
Because the RTF's analysis is based on existing evaluations, the impact estimates are backward looking. That is to say that the estimates are based on what has occurred in the past, which may not be indicative of the future residential water heating market. The average demand response impact that a program can expect during an event will largely be determined by the technology and efficiency of the underlying water heating systems enrolled. This mix of underlying technologies is expected to shift meaningful in the coming years as the market share of heat pump water heaters increases, and as heat pump technology progresses. Additionally, household usage of hot water (and coincident peak demand) has shifted over time, and may continue to shift, due to water saving measures, household behavioral changes, and other factors.
For heat pump water heaters specifically, most models have several operational modes for the household to choose from. One primary differentiation between these modes is their propensity and criteria to utilize electric resistance back-up heating. The RTF analysis has no data on the distribution of mode settings in regional households, or how different modes correlate to demand response impacts.
At the time of approval, PNNL was performing additional analysis on the data collected in BPA's CTA-2045 study to possibly generate more granular results based on heating zone, housing type, of other factors. Based on RTF guidance, these additional results may be incorporate into the analysis at a future date if appropriate.
Have questions? Please get in touch.
